IRIS Multimodel Suite

This is a set of modules with real world data using Intersystems IRIS Multimodel concepts.

Demo

You can see it working at: http://iris-multimodel-suite.eastus.cloudapp.azure.com/cryptocoins-market

Crypto Coins Market Graph

The crypto coins market isn't only BTC (Bitcoins), if we take a look at the biggest Crypto Coin Exchange (Binance) we can see plenty ways to trade crypto coins. And how IRIS Multimodel Pattern can help us to do it? In this particular case this application uses IRIS with 2 datamodels. 1st to ilustrate the link between market crypto coins and altcoins I am using the globals and the value stored is a json document like. Using this graph you can click at each symbol to trade using the binance home broker, but if you want, you can create your own home broker using all the features of python + Intersystems Iris.

Fish Species Database Using Globals

Ingesting data from fishbase we can store all relations between family, genus and species of fish. And by click or text we can query to see what we want + the fish picture. This is an interactive graph which can make easier the learning path of each specie. And depending on which data is ingested, is simple to render this graph.

http://iris-multimodel-suite.eastus.cloudapp.azure.com/blog-post

Blog Post in a key-value database model

My engine works in a way that I cant imagine more simple than that. To persist each post I create a global "^blog", with the subscript "post", and the next subscript is the post Id. In this global I put the content of the post and finish! Just this, nothing of create table, index, etc... using the values persisted I render the markdown into html and show on page.

^blog("post", "1") = "# post 1 content..."
^blog("post", "2") = "# post 2..."
^blog("post", "3") = "# post 3 markdown content..."

You can see the demo here:

Multimodel?

Almost market developers used to use one technology for each data model concept. With Intersystems Iris you can do it with the same systems. In a short time we can storage a graph data structured into Iris Globals, a document and a common SQL model. On crypto coins markets, I have put each cryptocoin into a global carrying it's name. And if the cryptocoin has any other cryptocoin which can be traded, you can easily create a link between then with a comma =)

How to persist data into IRIS Globals?

You can use plenty languages to do it e.g.: ObjectScript, Java, .Net, Python (my case). In this project I get all Binance data connect an Python application to the Exchange API and ingest all data using Native API in Python is just a matter of preference by the developer.

Tools

A set of classes in python using the IRIS Native API:

• irisdomestic: A class that I made to show one way I use the Native API extending the native api.

#has the same methods of irisnative + factory of irisglobal class

• irisglobalchart: A component to plot any global as a network graph chart.

• irisglobal: A class that I made to be filled as a Graph Data Structure and all recursive. So if you instatiate one irisglobal object all global data will be in memory in this object.

Imagine a global like this*

^covid19("countries", "us")=5000
^covid19("countries", "us", "newyork")=10
^covid19("countries", "brazil", )=100

With my class irisglobal in python you have just to instatiate to have access to all global nodes in memory and indexed as a dictionary.

obj_global = irisglobal("^covid19")
print(obj_global.subscripts["countries"].subscripts["brazil"].value)
100

Getting started

Prerequisites

• git
• docker and docker-compose adjust docker settings to up memory and cpu the AI demands more capacity
• access to a terminal in your environment

Installing

After cloning this repo open a terminal go to the iris-python-covid19 folder and type these commands:

git clone https://github.com/renatobanzai/iris-multimodel-suite.git

Running in linux and MacOS

docker-compose build
docker-compose up

Estimated time to up containers

1st time running will depend of your internet link to download the images and dependencies. If it last more than 15 minutes probably something goes wrong feel free to communicate here. After the 1st time running the next ones will perform better and take less then 2 minutes.

If is everything ok

After a while you can open your browser and go to the address:

• Main Menu: http://localhost:8050

Main Menu

The project has a main menu that points you to all the examples. Feel free to navigate.

• Cryptocoins Market Graph

You should look at IRIS Admin Portal

I'm using for now the USER namespace (todo: create my onw namespace)

http://localhost:52773
user: _SYSTEM
pass: SYS

Multimodels applied

Key-Value (Globals)

Cryptocoins Market are based on globals key-values, if you take a look at namespace USER you can see the global "^binance". I use the subscripts to connect the cryptocoins pointing which one can be traded by the other. To render I use a python module created by me: irisdomestic.py. You can see the code at _main_.py

# ingestion methods
def ingest_binance_markets():
# reading global methods

def update_cryptocoin_graph(global_text):

SQL

As fish families are not available by API, I use JDBC on python to query a table to get all families from the table fish.family in: _main_.py

def get_fishfamily():

Object

To record all visitors data I decided to use the object data model. Create the class on intersystems IRIS file: ./iris/src/banzai/visit.cls . The result can be viewed at the page /visit-log. I persists the data using the IRIS Native API at _main_.py:

# persising data
def save_page_visit(page_name):
    obj_irisdomestic.iris_native.classMethodValue("banzai.visit",
                                                  "SaveVisit",
                                                  request.remote_addr,
                                                  page_name,
                                                  datetime.datetime.now().timestamp()
                                                  )

My own model

I have used on science fish a combination of document and graph model database to store on globals more than on text. So inside the ^fish there is a dictionary with values that helps me to make the chart richer and more interactive (with url image, global address, and other data). This can be observed on methods:

# __main__.py
def update_science_fish_graph(global_text):
# iris_python_suite.py / irisdomestic class

def fill(self):

How does it work?

This is a python application using the IRIS service to persist and read data. I use globals to store raw data from JHU and plot it using Python community libraries. All code in ./app folder. Here some articles link to understant better the application:

• iris-python-suite-hitchhikers-guide-global-1
• using-python-represent-globals-network-chart

If you enjoyed this application please vote in iris-multimodel-suite!

https://openexchange.intersystems.com/contest/current